This report is primarily for explorative purposes and should not be used to draw any definitive conclusions about the data. Such conclusions should be drawn from rigorous statistical analyses in collaboration, if needed, with a biostatistician.
Statistical tests have been performed to help identify potentially interesting changes in the data. It should be noted that these tests are automated and assumptions have been made that may not be strictly correct depending on the behaviour of the data. For example, lipid data tend to be log-normal so, prior to statistical analyses, the data were log transformed to meet the assumptions of parametric tests. However, such transformation may not be suitable for all classes/species.
In no case, should the statistical results from the report be used in publications without further checks on the data.
A range of plots are shown below to help get an overall idea of how the data is behaving. Many of these plots show the same data but presented in slightly different ways.
Scroll down to view all of the plots or click on the links below to navigate directly to a plot type.
Pie charts
Box and whisker plots
Forest plots
Volcano plot
Heatmaps
The pie charts show the proportion of lipid types in the samples. Any that comprised < 1% of the total amount were grouped into an “other” category for the pie charts. The full set and values can be found in the table below.
The plots and table show total summed values across all replicates.
The box-and-whisker plots show 5 point summary statistics for the lipid class: the median, upper and lower quartile (box) and min/max values (whiskers). Points beyond whisker limits are outliers.
Each data point for the class is overlaid in the corresponding colour.
These plots share the same features as those above, but show the log2 transformed data.
T-tests were performed on log2 transformed data. Classes are annotated with a star if there was a significant difference between the conditions (adjusted p-value < 0.05).
The red point shows the ratio while the lines show the confidence interval. If the confidence interval does not span 0, the difference between the 2 conditions is likely to be significant. Red points to the left of the line indicate higher average levels of the lipid class in the Control condition, while points to the right of the line indicate higher average levels in the KO condition.
The y axis shows the –log10(padj); that is the negative of the log-transformed p-values, adjusted for multiple comparisons.
The x axis shows the magnitude of difference between the conditions for that lipid, expressed as a ratio.
Any points to the left of the grey dashed line have higher values in Control compared to KO, and vice versa for points to the right of the grey line.
Each point on the plot is a lipid subclass.
Hover over a point to see the name.
The dashed red line shows significance (adj p<0.05) i.e. any points above it represent a significant difference between the 2 conditions.
These heatmaps show log2 transformed data and are scaled by row, to allow comparison between samples.
This heatmap shows lipids with an adjusted p value < 0.05 from t-tests. Row labels are shown if there are < 70 rows.
Rows and columns have been clustered using default clustering methods in the pheatmap R package.
Rows ordered by lipid subclass, columns ordered by condition, no clustering.
This page shows 2 plots per lipid class; these show the lipid subclasses within each class.
The box-and-whisker plots show 5 point summary statistics for the lipid subclasses: the median, upper and lower quartile (box) and min/max values (whiskers). Points beyond whisker limits are outliers.
Each data point for the class is overlaid in the corresponding colour.
The plots on the right show log2 ratios between the 2 conditions, the red point shows the ratio while the lines show the confidence interval. If the confidence interval does not span over 0, the difference between the 2 conditions for that particular species is likely to be significant.
Subclasses that did not fit the assumptions for two-sample t-tests may not be shown in these plots.
Scroll down to view all of the lipid classes or select the blue links below to jump directly to a class.
This page shows individual data points for each lipid subclass.
p-value (uncorrected) : 0.0115
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.0929
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.287
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.00355
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.0481
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.191
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.658
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : NA
The data did not meet the criteria for performing a statistical test. There may have been too few non-zero values or the variance of a condition may have been too high.
p-value (uncorrected) : 0.0292
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.0389
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.0087
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.0429
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.0124
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.00399
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.000733
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.00197
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.239
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 3.25e-05
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.00394
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.000494
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.00337
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.61
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.245
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.00115
The data did not meet the criteria for performing a two sample t-test. One of the conditions had almost all 0 values, so a one-sample t-test was performed for the other condition. The results of this test may not be very robust and the plot should be viewed carefully to understand the data.
p-value (uncorrected) : 0.428
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.00499
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.00142
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.00375
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.0116
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.00913
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.0039
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.327
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.021
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.00247
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.186
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.0053
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.531
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.00342
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.279
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.779
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : NA
The data did not meet the criteria for performing a statistical test. There may have been too few non-zero values or the variance of a condition may have been too high.
p-value (uncorrected) : 0.00492
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.675
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.00172
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.00503
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : NA
The data did not meet the criteria for performing a statistical test. There may have been too few non-zero values or the variance of a condition may have been too high.
p-value (uncorrected) : 0.71
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.18
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.254
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.0723
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.0145
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : NA
The data did not meet the criteria for performing a statistical test. There may have been too few non-zero values or the variance of a condition may have been too high.
p-value (uncorrected) : 0.101
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 0.000513
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 8.9e-05
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : NA
The data did not meet the criteria for performing a statistical test. There may have been too few non-zero values or the variance of a condition may have been too high.
p-value (uncorrected) : 5.41e-05
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : 4.28e-05
Paired t-test performed on log2 transformed data. The samples were paired/matched.
p-value (uncorrected) : NA
The data did not meet the criteria for performing a statistical test. There may have been too few non-zero values or the variance of a condition may have been too high.
This report has been generated on the assumption that the type of experiment carried out was a matched design.
We have considered 2 types of experimental designs which we called independent and matched respectively. They are defined as follows:
Independent design: 2 or more conditions (e.g. genotype, treatment …) and in each several independent biological replicates (e.g. mouse, human …).
Matched design: aka repeated, dependent or paired. For this type of design we have considered 2 possibilities:
Design1: Two or more measures per animal/subject/petri dish (e.g. before/after treatment measures).
Design 2: Two or more independent experiments and within each experiment: same conditions (e.g. in each experiment 1 WT and 1 KO mice, the 2 animals are considered matched).
0 values have been removed.